1 Widner

Getting Things Done: Building Institutions in Challenging Contexts

Jennifer Widner Professor of Politics & International Affairs Princeton University

October 2012

For some time, now, economists have suggested that levels of income per capita will gradually converge across countries.1 We have seen this process start to happen in China and India where income per capita has risen. However, at the same time, we have watched another set of countries diverge. There is a big and persistent gap—and that gap is growing in a few places, especially in the poorer post-conflict countries.

It is now commonplace to attribute that gap to “institutions” and to observe that the norms and rules with which a community started are difficult to alter. Differences in executive constraint, property rights and rights enforcement, and other institutional features appear to shape trajectories as much or more than financing gaps, technology, or human capital. While some countries or communities acquired inclusive economic and political institutions that favored investment, trade, and innovation at some point in the past, others did not. Since that time, growth has helped to strengthen institutional capacity in the countries that found ways to limit executive abuse of power and encourage enterprise. At the same time, countries with more extractive institutions have found it difficult to change.2

It is this problem I want to address today. How can a community get from where it is now to a better set of institutions? Do people just have to wait until an event shatters past institutions and the deeply engrained interests and practices that accompany them? Or is it possible to offer some general maxims about how to make that transition?

A little over three years ago, Princeton’s Innovations for Successful Societies program decided to try to take on this challenge. Instead of starting with a general theory, we decided to go to the ground and study efforts to transform single agencies and scale-up or diffuse these changes. Public servants in some communities have found effective strategies for building the kinds of institutions that will promote growth and foster common interest states. What could they tell us, through their actions, about how to improve capacity, reshape norms, and build self-reinforcing incentives to sustain these changes? ISS chronicles the strategies employed, the obstacles encountered, and the tactics that helped or hindered success. It employs longitudinal case studies, as well as comparisons that explore the results of similar strategies deployed in matched and dissimilar

1 For example, see Kenneth L. Sokoloff and Stanley L. Engerman, “History Lessons: Institutions, Factor Endowments, and Paths of Development in the New World,” Journal of Economic Perspectives, 14, 3 (Summer 2000): 217-232. 2 For example, Daron Acemoglu and James A. Robinson. The Origins of Power, Prosperity, and Poverty: Why Nations Fail. New York: Crown Books, 2012. 2 settings, to build and test generalization. Although the program does not offer a neat general theory, it points attention in some promising directions. That’s what I want to talk about today.3

What We Know We actually do know how to transform some kinds of services or functions fairly fast, speeding delivery, reducing leakage or corruption, and improving quality. To be more precise, we know a lot about how to produce a single-agency turnaround when the services that agency produces are fairly standard and do not require extensive judgment or discretion, and where there are limited numbers of distribution points. Unlike teaching, health care, or policing, the services these agencies perform involve fairly repetitive actions where the criteria for acting one way or another are clear— for example, institutions that register land, issue identity documents or business licenses, or collect taxes.

One ISS project line has focused on such turnarounds. Our research teams have documented transformation strategies in this kind of agency in both post-conflict settings and countries with lower levels of internal strife and higher levels of capacity. The successful strategies share essentially the same features, everywhere. The reformers follow seven basic steps. They

 Trace the delivery chain to figure out where delays or other problems were taking place  Standardize procedures to clarify expectations and limit space for corruption  Set clear targets and advertised what members of the public should expect  Streamline procedures and space  Create a system to monitor action items and progress toward targets  Conduct informal internal audits—usually surprise visits  Mobilize social pressures to win compliance from staff members.

Although the elements seem to leap from the pages of a project management textbook, they appear to have practical origins and to have been rediscovered over and over. They were present in whole or in large part in every instance. In a few cases the reform leader brought in consultants— domestic or international—who helped organize the process, but reformers invented the steps themselves where consultants didn’t suggest them.

The story of South Africa’s revenue agency provides an example. In the late 1990s, after the transition to majority rule, the revenue agency faced several challenges. It had to diversify its ranks to include more black South Africans, but it also had to address slowness and low tax compliance. A team that included Pravin Gordhan, Jacques Meyer, Judy Parfitt, and several other agency officials sought to make big changes without layoffs. As a first step they worked with unions to develop a plan for moving forward, appointed new management trainees to work alongside the reform team, and then walked through the organization’s offices, tracking procedures and also spotting people who had the talent to lead reform. Using this information, they created change teams and then began to

3 Innovations for Successful Societies case studies, interviews, and cross-cutting analysis are available at http://www.princeton.edu/successfulsocieties 3 Widner streamline and automate many operations. They separated the front and back offices and re- engineered space to speed up lines and to limit personal contacts (which required some outreach to the public, whose members had often developed relationships with particular tax office officials). They made tax forms easier to complete and to read. They automated where they could, in order to spot evasion but also to help taxpayers correct mistakes. They shifted people no longer needed for processing forms to compliance work, analysis, and trouble-shooting. They took steps to focus the agency’s workers on outcomes and made sure each person knew exactly what his job was and how it contributed to the organization’s larger goals. They made it easier to pay taxes by creating “tax days” at companies and public libraries, where they would set up their computers and provide help in completing forms. Between 1998 and 2009, the South African Revenue Service dramatically improved tax compliance. The number of income-tax payers increased to 4.1 million from 2.6 million during the period.

The South African example points up some other common themes in these single-agency turnarounds. First, for the most part, the words “merit system” and “performance management” do not enter these stories. Although scholarly writing has often singled out merit systems as the source of success, any practitioner knows that performance appraisals and the other elements of contemporary merit systems are themselves very difficult to create. They require clear job descriptions, goals, and metrics, as well as extensive coordination across institutions—prerequisites that don’t exist in many countries. They often necessitate changes in the norms that govern inter- personal relationships. They are useful, but for many institutions and countries, they are not feasible short-term goals.

The South African Revenue Service had some of the ingredients of a merit system in place, but it motivated employees and built capacity largely as a result of the leadership of the change teams it assembled at the branch level, better matching between aptitudes and tasks, a shared sense of mission, and creation of a clear link between the organization’s activities and public appreciation. All of the reformers whose agency turnarounds ISS has chronicled have found a way to mobilize social pressures in order to help build performance, absent a clear merit system. For instance, they used simple devices weekly contests between work groups—or reporting out periods that helped people set standards. In most instances, the reformers rewarded groups, not individuals. The groups took care of creating the links to individual performance.

Second, although some of these agencies paid higher wages or offered bonuses to small groups or to a few high performers, many did not. Most of the single-agency turnarounds did not distort the wage bill, although a few did. Many of the single-agency transformations ISS has studied avoided use of compensation as an incentive, while others generated were able to use revenues from fees collected for services they provided to offer bonuses to the parts of the organization that exceeded targets. Higher wages or bonuses were not the main reason the turnaround occurred.

Third, bringing change to institutions that deliver services involving more discretion or services with many delivery points in remote areas poses extra challenges that make some of the 4 steps harder to execute and requires more attention to how to bring social pressures to bear. Nonetheless, a country that can improve many of its core economic services, which often have the characteristics of the turnarounds described here, and can improve urban services, where monitoring and social pressures are easier to mobilize than in rural areas, would accomplish much.

Why Better Institutions Don’t Emerge More Often If it is possible to identify the elements of successful strategies, why don’t we see improvements in institutional capacity everywhere? Many would-be reformers never “cracked the code” and followed the steps that brought success to a few. On this point, former Indonesian Finance Minister Sri Mulyani Indrawati reflects that after 1998, “Everyone [in Indonesia] talked about reform, but they really didn’t know how. Amazingly, not many people really understood how to translate ideas into reality.” Most reform leaders had no way to share experiences and learn from each other.

Distilling the elements of success in this subset of cases also risks giving the false impression that these institutional reforms were easy. They weren’t. The work involved in producing these turnarounds was hard and it was often dangerous. Idris Jala, CEO of Malaysia’s Performance Management and Delivery Unit, has noted about reform in his country: “The ideas were already there. People knew these things for a long time. But to move from idea to results—that’s where there are hurdles.” Nasir El-Rufai, Former Director General of Nigeria’s Bureau of Public Enterprises, has echoed that view and zeroed in on one of several important political obstacles. “The thing about reforms,” he observed, “is that they have a time lag. You start doing things that are difficult and the results don’t really come until a few years down the road—sometimes a decade, sometimes a generation.”4 Under these conditions, senior political leaders may hesitate because they fear failure.

What are some of the biggest impediments reform leaders face, and are there strategies that can help people get around these? I want to turn to what some of the other ISS case studies tell us about the challenges of institutional reform. My points fall under five headings: 1) overcoming the time inconsistency problem that Nasir El-Rufai noted, 2) lengthening the time-horizons of the people positioned to lead reform, 3) escaping spoiler traps, 4) improving focus and coordination at the top, and 5) creating reputational networks. This list is not all-inclusive, but these elements play an important role in the nearly 100 cases of institutional reform ISS has documented.

1. Overcoming Time Inconsistency in Reform

The benefits of reform accrue in the future, after hard decisions, changes in practice, or sacrifice. For example, a mayor who wants to build infrastructure to boost the economic potential of

4 Interviews conducted by Innovations for Successful Societies, accessible at the program’s website: http://www.princeton.edu/successfulsocieties October 2012. 5 Widner a city often has to turn to taxpayers or businesses first for revenue or cut the budget for other activities. Where there is a record of performance that engenders trust, the mayor may find it relatively easy to raise the funds, but in many places a trail of broken promises has created the opposite condition: a legacy of distrust that means citizens and businesses are often unwilling to provide support, figuring that they will never see anything for their contributions. This problem is more intense at the national level, where the services a reform leader wants to fix may seem remote from citizens’ daily lives.

Idris Jala’s adage in leading reform in Malaysia—“Big Results Fast”—captures the essence of the strategy most of the successful reformers ISS has interviewed use. To send a clear signal that “this time is different” and elicit enthusiasm from citizens and civil servants who have grown cynical, the first step is to produce rapid and significant improvement in a service people use frequently and can observe directly. Sudden reductions in processing times for documents people need on a regular basis, improvements in electricity or water, faster removal of trash, a drop in traffic congestion, sharp reductions in traffic police corruption or crime—all of these are candidates for demonstrating both resolve and the ability to deliver on promises. Publicizing the results is also a central part of this strategy. Effecting this kind of norm change is easiest in urban areas because of high population density and the focal points that cities afford. The projects chosen for this purpose must allow people to see a clear relationship between new tax monies raised or changes in practice and real results.

At the helm of Malaysia’s national effort to improve government performance, Jala worked with teams of public servants to identify a few key priorities, set big targets—stretch targets, he called them—define metrics, and report results in public. In other settings, such as Medellin (Colombia), Solo (Indonesia), Naga City (Philippines), Lagos (Nigeria), or Palermo (Italy), where reformers worked at the local level, similar strategies focused on restoring public spaces, relieving traffic congestion, or providing more regular water and power supply.

2. Lengthening time horizons

It takes time to build a new institution or to plan and execute a reform. Although the ISS case portfolio is not structured in a way that makes it possible to generalize reliably about the frequency of one sort of obstacle compared to another, in the cases we have documented, the inability to remain on post long enough to see a reform through is one of the biggest impediments to initiating and sustaining change. Public servants cannot learn about their own organizations, develop plans, build coalitions, renegotiate roles, or launch pilots if they aren’t on post long enough to shepherd the process or see a return on their investment of time and care. Short time horizons can also lead to stop-and-go policies that yield few concrete results and create the trail of broken promises that undermine trust or believability.

Short time horizons have several common sources. The character of political competition is one potential cause of trouble. In some societies, incumbent political parties and opposition candidates perceive common ground and work to preserve changes that are inclusive and benefit 6 large numbers of citizens. The notion of a “loyal opposition”—an opposition willing to set aside short-term partisan advantage in order to serve broader citizen interests—sometimes enables reform leaders to endure political transitions. By contrast, where partisanship is intense and laws permit, political parties often retire, reassign, or dismiss public servants whose accomplishments materialized during the tenure of the party previously in power. We don’t see many single-agency turnarounds in this second kind of context. No public servant is likely to take the risk that a successful effort to improve a public service might trigger the loss of a job on which his or her family depends. Political rules that limit ability to fire or retire senior civil servants without cause help lengthen time horizons among the people whose leadership is essential in institution building.

Similarly, where civil service posts rotate rapidly, there is simply no time to take the steps required to reform an institution and thus no incentive to do so. In some countries, rapid rotation is enshrined in civil service rules. In others, brokers receive payments from people for securing transfers or promotions and it is in their interest for jobs to turn over rapidly. These markets for public office, a kind of “governance trap,” are especially pernicious. Because the brokers often pay legislators a kickback in order to keep them from writing laws that penalize their actions, it is very hard to reverse the problem, once it gets started. An effort is under way in Indonesia, currently, to wrestle with this problem. In the mid-1980s, Robert Wade documented a similar challenge in India and offered a proposal that shares elements with the Indonesian response.5

The steps needed to create support for institutional change and to manage the process of reform can also easily consume more than one electoral term. Most elected officials have terms between three and five years along, although some are as short as one year. The South African revenue authority reform took several years to accomplish, as many major institutional changes do. Where term limits exist, it often takes a “tag-team”—two sequential, like-minded mayors or governors—to see the reform process through. In countries where there is strong party discipline, this kind of cooperation may prove possible, but in countries where parties are weaker and cannot influence their members, this approach is less common. The “tag team” form of cooperation has played an important role in Colombia’s cities, in Lagos State, Nigeria, where the first mayor or governor in the pair laid the groundwork and helped build the coalitions, while the second managed changes in process and accepted the continuing informal assistance of the predecessor in maintaining ties with important constituencies. Bangalore was the site of an alternative strategy for coping with a mismatch between electoral terms and the needs of a reform agenda. There, a committee of civic groups provided the anchor for the reform process and required continuing reports from departments and a continuing process of target-setting, no matter who formally held office.

Finally, the structure of political incentives can create short time horizons by creating instability in the capacity to launch and sustain reform, instead of the reformer’s job status. For example, in Papua New Guinea, decentralization vested extensive powers to deliver services people wanted in local governments but gave the parliament considerable discretion over the size of block

5 Robert Wade. “The Market for Public Office: Why the Indian State is Not Better at Development,” World Development, 13, 4 (1988): 467-97. 7 Widner grants from the central government. Gradually parliamentarians realized the local officials could use the ability to deliver services as a campaign advantage and win election to national office by building on their experience, while incumbents in parliament had no comparable authority over the services most citizens valued. The parliament did not oust officials from their posts, but they did cut off some of the funding to local governments in order to preserve political advantage.

3. Escaping Spoiler Traps

Improving institutional performance almost always means making changes that cause an important person, family, or faction to accept some losses. A crisis or an unwilling of public anger can create some political space to promote change and enhance service delivery, but after a time, the people who earlier gained from dysfunction seek to revert to old ways. For example, creating standard operating procedures and streamlining—two common elements of single-agency turnarounds—frequently (and deliberately) reduce opportunities to divert resources to private uses. Unless the economy is growing and there are ways to replace lost income, licit or illicit, a backlash easily develops. Removing a former combatant from a leadership post in order to improve performance may even rekindle a conflict. This kind of “governance trap” emerges when a few well- positioned people who benefit from the status quo are able to block changes that promise improvement for the many. Spoiler traps often block institutional change.

Forging an alliance with the public. Malaysia’s Idris Jala has articulated the approach reformers often stress when asked about how to address a spoiler trap: Make the public the monitor and enforcer; build a coalition with the public. The risk of inciting public ire can keep potential spoilers from trying to have their way. In Malaysia, Jala invited people to come to town halls to see the reform targets his office had negotiated with different parts of the government. “By publishing and declaring it like that you have no choice. It is like a salesman’s nightmare. You are committed to deliver.”

Successful reformers in other countries have done the same, sometimes adding dimensions Idris Jala does not mention. They have created new sources of pressure on politicians and civil servants by the strategic use of transparency, public meetings, press engagement, student involvement, and media campaigns to shift public norms. As pressure mounts, they reduce the spoilers’ access to resources, knowing that most won’t react immediately because of the heightened public scrutiny. They use this political space to draw followers away from the spoilers and they try to build the base for future public demand by delivering concrete improvements in public goods fast.

In Indonesia, in 1998, public anger against corruption gave rise to “Reformasi,” a movement for institutional and political change. As part of the response, the president and other leaders created a powerful anti-corruption commission to take on some of the spoilers. In the initial years, the commission had political space to act because reformist ideas had captured the public imagination. As the spirit of “reformasi” started to wear off, however, the commission encountered push back 8 from the legislators, some of whom were part of the initial problem the public movement sought to address. In this instance the commission had the foresight to cultivate sustained demand among young people and key sectors of society that it could later mobilize to limit the “counter revolution.” It worked with Indonesia’s many universities to sensitize students to the issues and engage them in monitoring corruption. Although the story is not over, the effort to build a continuing social conflict prevented some significant reductions in effectiveness. (A two-part case appears on the ISS website.)

Rapid Results & Re-Setting Norms In lower capacity settings, the Idris Jala strategy is harder to pursue. Public monitoring and pressure is often harder to build, and senior officials or potential spoilers know that it will be very difficult for reformers to sustain civic engagement. In these settings, small “under the radar” projects that gradually reshape public expectations and political incentives may be helpful.

These initiatives assume many different forms, but they have several elements in common. First they engage small groups of public servants, usually at the middle and lower levels, in conversations about the kinds of service improvements people want. With some coaching, each group chooses a project that will deliver a result within 100 days. Usually these are relatively small scale, but they include objectives like improving license processing speeds at a branch office, overcoming a barrier to fertilizer use, vaccinating a given number of children, etc. The group has to figure out how to meet the goal. Its members get a little bit of coaching and can ask for advice. The projects are a way to focus attention and teach basic project management. At the end of the 100-day period, each group must explain what it did and what it achieved at a public meeting. The higher performing groups win a prize. In most instances higher officials and local politicians are invited to the meeting, along with members of the press. The coaches ask the politicians and senior officials to sign a document endorsing the gains achieved. Press coverage gives the officials and politicians a pat on the back, enabling them to take some credit. It also sets up public scrutiny, ensuring that the press and civic groups hear the politicians promise to support similar initiatives in the future. A second round of 100-day projects then starts.

These approaches ensure that there is some progress in delivering concrete results and re- setting expectations even if there is political turbulence at the top. The projects are not so large that spoilers are often an issue, but the engagement of politicians in a public forum gradually reshapes incentives—or so the theory goes. (Most of these programs have not been in existence so long that the effect on incentive structures is wholly apparent.)

External agencies of restraint. The rapid results interventions are helpful in generating some concrete results, re-setting leadership norms, and slowly building the base for civic engagement. However, they don’t touch the tougher and more immediate spoiler problems that often exist at the end of a conflict. Leverage in these instances usually requires some type of outside engagement to undermine the power of potential spoilers. The outsiders may be multilateral organizations, regional organizations, or more rarely, banks. 9 Widner

In 2006, Liberia’s only functioning seaport was a quagmire, riddled by corruption, cargo theft, and a glut of untrained workers. These problems combined to slow the delivery of relief supplies that were badly needed after a 14-year civil war, which had ended three years earlier. A battle site, the Freeport of Monrovia suffered from war damage and years of neglect. It was in danger of shutting down completely. The responsibility to upgrade the infrastructure and improve management lay with the National Port Authority (NPA), a state-owned enterprise. From 2006 through 2011, Togba Ngangana, George Tubman and Matilda Parker, successive managing directors at the NPA, enacted a series of reforms to restore the authority and port operations. The Liberian government and outside donors agreed to hire internationally recruited financial controllers to work with NPA directors, under the terms of GEMAP—the Governance and Economic Management Assistance Program. Liberians and internationals sat together at the same table to take decisions. The program also placed internationally recruited controllers and technical experts and assigned them key roles in the port’s management. Further, the United Nations Mission in Liberia (UNMIL) temporarily took control in 2005 and removed 370 members of the port security staff—sending some for training and laying off others. This action also helped reduce a potentially serious spoiler problem.

4. Improving coordination at the top

A fourth common obstacle to building, sustaining, and diffusing institutional capacity is the frequent inability of the cabinet office or the president’s office to coordinate across ministries in the event of a bottleneck or a problem. In many of the successful cases of single-agency turnarounds, the coordination demands are low, but in a few, cooperation from port authorities or the treasury or another department is essential. In those instances, a well-functioning cabinet office is essential, yet cabinet performance is often weak, especially in post-conflict settings.

Heads of government encounter a number of obstacles in trying to support institution building. Some struggle to communicate their priorities to cabinet members and to the civil servants in ministries and agencies on the frontlines of policy implementation. Even if the priorities are well communicated, ministries and agencies might pursue their own long-held institutional goals and relegate the president’s or prime minister’s priorities to the backburner. Follow-up is essential. The sometimes thinly staffed offices of heads of state or heads of government sometimes lack a way to focus on priorities and to obtain quality policy documents that can help principals make thoughtful decisions.

Some governments have launched “delivery units” to overcome these organizational barriers. In creating a delivery unit, a head of government articulates a small set of priority goals to stand out among government projects. To ensure ministries indeed focus on priorities, delivery units set up tracking systems communicate regularly with relevant ministries on behalf of heads of government to inquire about progress and delays. Delivery units offer assistance by helping line ministries map out the delivery chain, or the steps involved from high-level policy formulation to frontline delivery, with an eye towards removing obstructions and inefficiencies. Delivery unit staff may also help prepare or vet policy papers to ensure that the cabinet members have a clear sense of the choices and 10 options they face. The units brief the heads of government regularly, who, in turn, commit to use the power of their offices to break through delivery bottlenecks. By doggedly pursuing line ministries, delivery units make sure the agenda of the heads of government were reflected in the plans, actions, and achievements of their administrations.

Depending on how it is organized and managed, a delivery unit can facilitate coordination and make the jobs of ministers, as well as the head of state, more productive. Focusing on a few priorities is key. Expanded mandates run the risk of diluting the time and energy of the delivery unit leadership and of the heads of government. Finding the right people to staff the policy analysis positions is also important; the wrong people in these positions can cause discord.

5. Creating reputational networks

A fifth broad source of variation in ability to launch, diffuse, and sustain reform is the number of people available to manage these processes effectively—and the ability to identify those who have the right set of talents and experience. Idris Jala had business experience at Malaysia’s national airline before he helped create the country’s ministerial performance management system. Malaysia had many mid-level public servants it could coach to help out. Other countries have also tapped “Mr. Fix-Its” from the private sector as well as within the ranks of government. In Malaysia, South Africa, many of India’s states, Brazil, and other countries enough knowledge about the past accomplishments of particular individuals circulated that it was possible to identify reform leaders who could command broad respect and whose efforts would enjoy some protection from partisan sparring, or at least a honeymoon up front.

Other countries may be less fortunate. After a political transition, it takes time for reputations to emerge, especially where people rotate rapidly among posts. A new government won’t always know who has the skill to direct reform. People are untried. Reputations are inchoate. Information that would facilitate the identification of the rights skills doesn’t circulate well. A country with a lively private sector can sometimes borrow talent seconded from business, but the private sector has often shrunk during war and suffers the same problems. Attracting talent from the diaspora can prove costly if a country must boost wage rates, housing and school availability.

What are the options available to help identify people who could be recruited to lead reform effectively? This kind of “elite recruitment” sometimes emerges from reform strategies that start with several relatively small pilot projects, which enable self-made reform leaders to emerge. The leaders of the pilots that succeed in the first round go on to a second round of more complex pilots and so on until the “Mr. Fix-Its” come into clear focus. The advantage of this strategy is that it also provides experiential training and it frequently generates new ideas or innovations. When Malaysia sought to improve the performance of its land registries, at least part of its program pursued this approach. 11 Widner

Conclusion

In Why States Fail, Daron Acemoglu and James Robinson suggest that it is impossible to theorize about strategies for breaking out of path-dependent institutional trajectories. If rules shape behavior, create vested interests, and condition the possibilities of change in myriad other ways, then only a shock or “critical juncture” can open up a country to real reform and to a dramatically new institutional paradigm. By definition, these are periods when interests shift and bargaining positions become hard to assess. Change is necessarily idiosyncratic and unpredictable. We are better at understanding stasis than change.

This paper asks whether the process of institutional change is really as unpredictable as Acemoglu and Robinson imply. The nearly 2000 reformer leaders Innovations for Successful Societies has interviewed in the past four years think they have strategies that have generalizable elements. They think they can “crack the code.” The stories will not be exactly the same everywhere because conditions and the strategic structure of reform will vary across settings, but there are many common elements. By sharing experiences reformers are able to hone their sense of the best tactics and sequences for their own societies—and offer scholars and policy makers the raw material for generating broader insight.